Klasifikasi Gender Berdasarkan Fingerprint Menggunakan Metode Naive Bayes Classifier

Abstract

No two fingerprints are identical, as everyone has their own characteristics. The most fundamental problem lies in the results of the fingerprint image, typically due to inconsistencies in the emphasis of the fingerprint and the position of the fingerprint, resulting in inconsistencies in the thickness of the black line and shifting positions, which negatively impact the overall performance of the system. To solve this issue, research is required on a classifier that assumes all attributes exist independently. The NBC (Naïve Bayes Classifier) is a classifier based on the assumption that all attributes are independent. The NBC method for gender classification based on fingerprints consists of three steps. The initial step is to evaluate the quality of the image to be processed. This is demonstrated by the consistency of the grayscale values, which are not skewed when converted to a binary image. The second is the selection of data that exhibits no data deviation, which also leads to errors in the classification procedure that follows. With the existence of machine learning, class-based measurement formulations can be acquired through training. Even with unbalanced data, it is preferable to use NBC for classification purposes.